A comprehensive study on membrane fouling in submerged membrane bioreactors operated under different

Available online at

SeparationandPuri cationTechnology

59 (2008) 91–100

Acomprehensivestudyonmembranefoulinginsubmergedmembrane

bioreactorsoperatedunderdifferentaerationintensities

FangangMeng ,FenglinYang,BaoqiangShi,HanminZhang

KeyLaboratoryofIndustrialEcologyandEnvironmentalEngineering,MOE,SchoolofEnvironmentalandBiologicalScienceandTechnology,

DalianUniversityofTechnology,Dalian116024,PRChina

Received8December2006;receivedinrevisedform22May2007;accepted29May2007

Abstract

Inthispaper,membranefoulinginthreeparallelMBRsoperatedunderdifferentaerationintensities(150,400and800L/h)wasstudiedtohaveabetterunderstandingofthemembranefoulingmechanism.Theimpactofaerationonmembranefoulingwasinterpretedfromtwoaspects:evolutionofbiomasscharacteristicsandformationmechanismofthecakelayer.Theresultsshowedthateithersmallorlargeaerationintensityhadanegativein uenceonmembranepermeability.Thelargeaerationintensityresultedinaseverebreakupofsludge ocs,andpromotedthereleaseofcolloidsandsolutesfromthemicrobial ocstothebulksolution.Thesludgesupernatantwouldbecomeheterogeneousastheaerationintensityincreased.AstheMBRoperatedunderhighaerationintensityof800L/h,colloidsandsolutesbecamethemajorfoulants.Inaddition,thebacktransportmechanismofmembranefoulantsinthethreeMBRswasdifferentfromeachother.Aerationhadapositiveeffectoncakelayerremoval,butporeblockingbecamesevereasaerationintensityincreasedto800L/h.Themaincomponentsoforganicmattersinthemembranefoulantswereidenti edasproteins,polysaccharidematerialsandlipidsbytheFouriertransforminfraredspectroscopy(FTIR).© 2007 Published by Elsevier B.V.

Keywords:Membranebioreactor;Membranefouling;Aeration;Sludgecharacteristics;Cakelayer

1.Introduction

Membranebioreactorisabiologicalwastewatertreatmentprocessthatusesmembranetoreplacethegravitationalsettlingoftheconventionalactivatedsludgeprocessforthesolid–liquidseparationofsludgesuspension[1].MBRs,inwhichbiomassisstrictlyseparatedbyamembrane,offerseveraladvantagesovertheconventionalactivatedsludgeprocess,includingahigherbiomassconcentration,reducedfootprint,lowsludgeproduc-tionandbetterpermeatequality[2].AmajorobstaclefortheapplicationofMBRsistherapiddeclineofthepermeation uxasaresultofmembranefouling[3–8].

The rstgenerationofMBRswassidestreamorcross- owsystemswiththemembranemoduleplacesinarecirculationloopexternaltothebioreactor.Theuseofrecirculationloopsleadstoincreasedenergycosts.Inaddition,thehighshearstressesinthetubesandrecirculationpumpscancontributetothedestructionofbio ocsandthishasbeenlinkedtoaloss

Correspondingauthor.

E-mailaddress:fgmeng80@(F.Meng).

ofbiologicalactivity[9].Toovercometheselimits,thesub-mergedMBRsweredevelopedandpopularlyusedinwastewatertreatment[10].InasubmergedMBR,shearstressiscreatedbyaeration,whichnotonlyprovidesoxygentothebiomass,butalsomaintainsthesolidsinsuspensionandscoursthemembranesurfacetoalleviatemembranefouling.Thenormalprocessofaerationcanbeusedtogenerateashearstressonthemem-branesurfacewithoutrequiringarecirculationpump.But,ithasbeenfoundthatmorethan80%energyconsumptionwasforaeration[11].Uedaetal.[12,13]examinedtheeffectofaera-tiononcakeremovalandsuctionpressureusingapilot-scalesubmergedMBRandconcludedthataerationwasasigni -cantfactorgoverningthe ltrationconditions.Previousworks[14,15]alsoshowedthatthecake-removingef ciencyofaera-tiondidnotincreaseproportionallywiththeincreaseintheair owrateandthattheair owratehadanoptimumvaluefromthecake-removingpointofview.

Ahighaerationratecertainlycanreducesludgeattachmenttothemembrane,butitalsohassigni cantin uenceonthebiomasscharacteristics.Mostofthepreviousliteraturesfocusedoninvestigatingthein uenceofaerationintensitiesonmem-branepermeabilityandbiomasscharacteristics,butthereislittle

1383-5866/$–seefrontmatter© 2007 Published by Elsevier B.V.doi:10.1016/j.seppur.2007.05.040

92F.Mengetal./SeparationandPuri cationTechnology 59 (2008) 91–100

informationonhowtospecifytheimpactsofaerationintensityontheformationofmembranefoulants.Infact,thereshouldbeadirectrelationbetweenaerationintensitiesandtheformationofmembranefoulants.Theaerationintensityisexpectedtohaveaverycomplexin uenceonMBRperformance.

Inthispaper,threesubmergedMBRsunderdifferentaera-tionintensitieswereoperatedforabout50daystoinvestigatethein uencemechanismofaerationintensityonmembranefouling.Inthewholetests,themembranepermeate uxwasmeasuredtostudythemembranefoulingbehaviorunderdiffer-entaerationintensities.Thesludgeparticlesizedistributionsofsludgesuspensionandsludgesupernatant,solubleCOD,col-loidalCODandEPSwereanalyzedtocharacterizetheeffectofaerationintensitiesonbiomasscharacteristics.Thefoulinglayerformationmechanismwasexaminedtodescribethedepo-sition/adsorptionmechanismofmembranefoulants.Thefoulingcakelayeronthemembranesurfacewasanalyzedbasedonresistanceanalysis.Thedepositionofbiopolymersonthemem-branesurfacewascharacterizedbyFouriertransforminfraredspectroscopy(FTIR).2.Materialsandmethods2.1.OperationofMBRs

AsshowninFig.1,theexperimentalsystembasicallycon-sistedofthreeactivatedsludgebioreactor(MBR-A,MBR-BandMBR-C).Ineachbioreactoramembranemodulewassub-merged.Theeffectivevolumeofthebioreactorwas12L.EachMBRwasasubmergedhollow bermembranemodulemadeofpolyethylenethathadatotalareaof0.1m2andanormalporesizeof0.1 m(DAIKI,Japan).PriortotheMBRsoperation,thesludgewasacclimatizedbyotherMBRs.TheMLSSconcentra-tionofeachactivatedsludgesuspensionwasadjustedtoabout6000mg/Lwithwaterpriortothemembrane ltration.Thecom-positionofthefeedwaterwasasfollows:sucrose(300mg/L),urea(78mg/L)anddipotassiumhydrogenphosphate(38mg/L)wereusedasthemainfeedforactivatedsludge,andcalciumchloridewasappliedasmineral.SodiumbicarbonatewasusedasabuffertoadjustthemixedliquorpHtoabout7.0.Thetemper-atureofthemixedliquorwascontrolledat25.0 Cwithelectricheaters.Thehydraulicretentiontime(HRT)rangedfrom10to12h,thesludgeretentiontime(SRT)wassetat30days.The

aerationintensitiesforMBR-A,MBR-BandMBR-Cwere150,400and800L/h,respectively.Theaveragedissolvedoxygenconcentrationsofthesethreeaerationintensitieswere3.21,4.76and6.50mg/L,respectively.

TheMBRcanbeoperatedintwomodes[16]:constant uxandconstanttransmembranepressure(TMP).WithrespecttotherealoperationofMBRsforwastewatertreatment,constant uxispreferabletoconstantTMP.ThemodeofconstantTMPissuitableforthestudyofmembranefoulingbecauseitcanprovidemoreinformationonmembranefouling.EventhoughconstantTMPisnotpreferable,therearestillmanyreportsabouttheapplicationofconstantTMPforlong-termwastewatertreatment[17–20].Inalltheseliteratures,theMBRswereoperatedwithagravitational ltrationmode,whichgeneratedaconstantTMP.Thegravitational ltrationmodehadlowenergyconsumptionandwascost-effectivetobuild[18].Themajorobjectiveofourworkistoanalyzethemembranefoulingbehavior,butnotthetreatmentperformanceofMBRsystems.Therefore,thethreeparallelMBRswerealsooperatedwithgravitational ltrationmodeorconstantTMPmode.

ThroughouttheoperationofthethreeMBRs,themembranemodulewasdrivencontinuouslywithaconstantlow-pressure,TMP=3.97kPa,whichwasinducedbyawaterheaddrop( Z=40cm).InconstantTMPmode,themembrane uxwilldeclineduringmembrane ltrationasaresultofmembranefoul-ing.Whenthe uxwassmallerthan6L/m2h,themembranemodulesweretakenoutandcleanedby ushingwithtapwatertoremovethefoulingcakeonthemembranesurface.Thus,theextentofmembranefoulingdegreecouldbeexpressedbythefrequencyof ushing.2.2.EPSanalysis

TheextractionofboundEPSwasbasedonacationionexchangeresin(CER,Dowex-Naform)method[21]:300mLsludgesuspensionwastakenandcentrifugedat2000×gfor15minat4 C.Thesludgepelletswereresuspendedtotheiroriginalvolumeusingabufferconsistingof2mMNa3PO4,4mMNaH2PO4,9mMNaCland1mMKClatpH7.Then,thesludgewastransferredtoanextractionbeakerwithbaf esandtheCER(80g/g-MLSS)added.Thesuspensionwasstirredfortheselectedstirringintensity(900rpm)andextractiontime(1.5h)at4 C.TheselectedEPSwasharvestedby

centrifu-

Fig.1.Schematicofthesubmergedmembranebioreactors:1,feedtank;2,balancebox;3,bioreactor;4,membranemodule;5,electricheater;6,airpump;7,air owmeter.

F.Mengetal./SeparationandPuri cationTechnology 59 (2008) 91–10093

gationofasampleoftheCER/sludgesuspensionfor1minat12,000×ginordertoremovetheCER.Thesupernatantwascentrifugedtwicefor15minat12,000×gat4 Cinordertoremoveremaining occomponents.BoundEPSwasobtainedafter lteringthesupernatantthrougha0.22 mmembrane lter.TheboundEPSwasnormalizedasthesumofcarbohydrateandprotein,whichwereanalyzedusingphenol/sulfuric-acidmethodandfolinmethod[22],respectively.

2.3.SolubleCODandcolloidalCODanalysis

SolubleCODandcolloidalCODweremeasuredinordertodeterminewhichcomponentsinthesludgesuspensionweremainlyresponsibleforthe uxreduction[23].SupernatantCODwasdeterminedaftercentrifugingthemixedliquorfor2minat3000×g.SolubleCODwasobtainedafter lteringthesuper-natantthrougha0.22 mmembrane lter.ThecolloidalCODwasobtainedbysubtractingthesolubleCODfromthesuper-natantCOD.2.4.FTIRanalysis

Thefouledmembranemodulewastakenoutfromthebiore-actorand ushedwithpurewaterastheoperationoftheMBRswasterminated.About200mLwashedliquidwastakenandplacedinadryerat105 Cfor24htoobtaindryfoulants.AFTIRspectrometer(EQUINOX55,Bruker,Germany)wasusedtocharacterizethemajorfunctionalgroupsoforganicmattersinmembranefoulants.KBrpelletscontaining0.50%(drypowder)ofthesamplewaspreparedandexaminedintheFTIRspec-trophotometer.Thespectrumwascalculatedfromtheaverage of256scansoverthewavenumberrangingfrom4000to400cm1ataresolutionof4cm 1.2.5.Particlesizeanalysis

Thesludgeparticlesizedistributionsofsludgesuspensionweredeterminedbyfocusedbeamre ectancemeasurement(FBRM)(ModelM400L,Lasentec,Redmond,USA).Thepar-ticlesizedistributionofthesludgesupernatantwasmeasuredusingaMarlverncounter(Zeta100,UK).2.6.Evaluationof ltrationresistance

Membraneresistancewasevaluatedbytheresistance-in-seriesmodelasfollows:Rt=Rm+Rp+Rc=

TMP(1)

Theexperimentalproceduretogeteachresistancevaluewasasfollows[24–26]:(1)theresistanceofmembrane(Rm)wasesti-matedbymeasuringthewater uxofde-ionized(DI)water;(2)thetotalresistance(Rt)wasevaluatedbythe nal uxofsludgewastewatermicro ltration;(3)themembranesurfacewasthen ushedwithwaterandcleanedwithaspongetoremovethefoulingcakelayer.Afterthat,theDIwater uxwasmeasured

againtogettheresistanceofRm+Rp.Theporeblockingresis-tance(Rp)wascalculatedfromsteps(1)and(3),andthecakeresistance(Rc)obtainedfrom(2)and(3).

Cakeresistanceisrelatedsigni cantlytocakespeci cresis-tanceandcakemass:Rc=αmc

(2)

wheremcisthedrycakemassandαisthespeci cresistanceperunitcakemass,whichvarieswiththebulkmatrixpropertiesandTMP.2.7.Others

Dissolvedoxygenconcentration(DO)wasmeasuredbyaDOmeter(55/12FT,YSICorporation,USA).Themixedliquidofsuspendedsolids(MLSS)concentrationwasevaluatedbyStandardMethods[27].3.Resultsanddiscussion

3.1.Behaviorofmembranepermeation

Theevolutionofpermeate uxduringthemembrane ltra-tionofactivatedsludgeispresentedinFig.2a.Itisimportanttonotethathigheraerationintensityresultedinalowerfoulingrateintheinitial ltrationtime.Itiswellknownthatcakelayeronthemembranesurfaceisthemainfactorthatresultsinmem-branefouling,andtherearetwooppositeactionsthatregulatetherateofcakelayerformation:permeationdrag,whichisgen-eratedbypermeate ux,increasedwithoperationTMP,andbacktransport,consistedofBrowniandiffusion,inertialliftandshear-induceddiffusion[28].Thehigheraerationintensityinducedahighershearforce,andremovedthefoulingcakelayerfromthemembranesurface.Therefore,thefoulingratedecreasedwithincreasingaerationintensitiesintheinitial ltrationtime.Obvi-ously,thecurveofFig.2acanberoughlyseparatedintotwophases:from0thto400thhourasphaseI,from400thto1100thhourasphaseII.InphaseI,themembranefoulinginMBR-AandMBR-Bwasmoresevere,indicatingthattheaerationintensitymayhavesomenegativeimpactsonbiomasscharacteristics.InphaseII,themembranepermeationofthethreeMBRsreachedarelativelysteadyvalue,suggestingthatthesludgesuspensionineachMBRhadbeenacclimatized.

Itcanbeseenthatthepermeate uxofMBR-AandMBR-Cdeclinedrapidlyafter10h ltration.ForMBR-A,itmayberesultedfromtheformationoffoulingcakelayerduetotheloweraerationintensity.TherewasahigheraerationintensityinMBR-C,butthepermeate uxalsodeclinedrapidly.Thepermeate uxofMBR-Breachedasteadyvalueafter100h ltration,suggestingthattheformationofafoulinglayerthatisactingasa“dynamicmembrane”withlowerpermeabilitythantheoriginal ltrationmembrane[29].Itshowsthattherateofparticleconvectiontowardsthemembranesurfaceisbalancedbytherateofbacktransport.Therefore,forMBR-B,severemembranefoulingshouldnotoccurwithrespecttotime.Assoonasadynamicmembraneformed,themembranefoulantssuch

94F.Mengetal./SeparationandPuri cationTechnology

59 (2008) 91–100

InthewholeoperationofthethreeMBRs,thefoulingdegreeofthethreemembranemodulesmainlyresultedfromtwofactors:differenceofshearforceandchangeofbiomasscharac-teristics.Intheinitial4h,thechangeofbiomasscharacteristicscouldbeignored,andthemembranepermeate uxwasbasi-callyin uencedbytheshearforce.Therefore,thedirectimpactofaerationintensitiesonmembranefoulingcouldbeobtainedbycomparingthepermeate uxofthethreeMBRsintheinitial4h,thedataareshowninFig.2b.Theaerationintensityhasapositiveeffectonmembranepermeability,suggestingthattheshearforcegeneratedbyairbubblescaneffectivelyremovethefoulantsdepositedonthemembranesurface.

Besidestheaerationintensityeffectonmembranefoulants,itmayhavesomeeffectonbiomasscharacteristicsalsosinceMBRsystemincludeslivingmicroorganismsandtheirmetabo-lites.Thebiomasscharacteristicswouldinturnhavesigni cantimpactsonmembranefoulingduringmembrane ltrationofsludgesuspension[31–34].Inordertoinvestigatethein u-enceofbiomasscharacteristicsonmembranefouling,short-termmembrane ltrationtestswereperformedassoonasthelong-termexperimentwasterminated.Intheshort-term ltrationtests,thesameaerationintensity,150L/h,wasadoptedforMBR-A,MBR-BandMBR-Cinordertoexcludetheeffectofshearforceonmembrane ux.Thepermeate uxofMBR-AandMBR-Bdecreasedslowlyandhadasimilardecreasetendency,Fig.2c.FromFig.2c,itcanbeseenthatMBR-Chadadramaticmembranefoulingproblem.Theseresultsshowedthattoohighaerationintensityaffectedthebiomasscharacteristicsde nitely,whichmayleadtotoomuchreleaseofEPSandthebreakageofsludge ocs.

3.2.Evolutionofbiomasscharacteristics

3.2.1.Particlesizedistributionsofsludgesuspension

Thesludgeparticlesizewasmeasuredafterthesludgesus-pensionhadbeenacclimatized.TheparticlesizedistributionsmeasuredbytheFBRMsystemcouldrevealthesizedistribu-tionsofthesludgesuspensionparticlesintheMBRs.StatisticresultsofthesludgeparticledistributionaresummarizedinTable1.Itcanbeeasilyseenthatthesizeofthesludgepar-ticlesvariedinarangeof3–450 m,andmorethan70%ofthesludgeparticleshadasizerangingfrom10to100 m.Thepeakpoints,indicatingthelargestparticlesizedistribution,andthemeansizeinthepro lesweregivenintheorderofMBR-A>MBR-B>MBR-C.Themeanparticlesizewascalculatedonthebasisofnumberofparticles.

Itwasreportedthatthesludgeparticlesthathaveasizesmallerthan50 mwouldaffectthemembranepermeationsig-

parisonofmembranepermeate ux:(a)evolutionofthemembranepermeate uxinthelong-term ltrationtests,(b)evolutionofmembraneper-meate uxintheinitial4hshort-term ltrationand(c)short-termmembrane ltrationtestsunderthesameaerationintensity(150L/h)afterthelong-termtests.

asEPS,solubleorganics,colloidalparticlesandsoon,couldberejectedorbiodegradedbythedynamicmembranecomposedoflivingmicroorganisms[30].Thus,thefoulantshavefewerchancestodepositonthemembranesurface.

Table1

StatisticresultsofsludgeparticlesizedistributionsinMBR-A(150L/h),MBR-B(400L/h)andMBR-C(800L/h)

Sludgeparticlesize( m)Mean

MBR-AMBR-BMBR-C

484130

Range3–4643–3983–398

Peak937454

Particlesizedistribution(%)<10 m0.1260.2260.826

10–50 m19.50426.68947.258

50–100 m49.60049.73639.531

100–200 m29.31622.11910.373

>200 m1.4561.2032.001

F.Mengetal./SeparationandPuri cationTechnology 59 (2008) 91–10095

ni cantly[35].Wenotedthatonly20%ofthesludgeparticleshaveasizesmallerthan50 minMBR-A,whereasmorethan48%ofthesludgeparticlesdistributedasizerangefrom0to50 minMBR-C(Table1).InMBR-C,thehighaerationinten-sitywasthemainfactorcausingtheformationofsmallparticles.FromTable1,italsocanbeseenthattherewere0.826%parti-clesthathaveasizesmallerthan10 minMBR-C.However,therewereonly0.126%and0.226%particlesthathaveasizesmallerthan10 minMBR-AandMBR-B.Althoughthehigheraerationintensitycouldinduceaneffectivebacktransport,thesmallparticlesinsludgesuspensionhadastrongtendencytodepositonthemembranesurface.Thehigheraerationintensitygeneratedastrongershearstress,andthenresultedinasevere ocbreakage.Thebreakageofthesludge ocscertainlyduetoerosionstrengthsortoaruptureofthenetworkofpolysaccha-ride brilswhichisthesupportofthedifferentcompoundsandparticularlythecells[36].

3.2.2.Particlesizedistributionsofsludgesupernatant

After30minsettlement,thesupernatantinthesludgesus-pensionwassampledanditsparticlesizedistributionwasmeasured.WithrespecttothesupernatantinMBR-A,therewasasharppeakat60nm(Fig.3a),suggestingthatmost

ofthesmallparticlesorsolutesinthesupernatantdistributedinthesizeof60nm.Thisresultalsoindicatesthatitwasarelativehomogeneoussystem.Withrespecttothesludgesuper-natantinMBR-B,thereweretwosigni cantpeaks:200and800nm,suggestingthereweretwoclassesofparticlesormacro-moleculesolutesinthissupernatant.InFig.3c,therewerethreesigni cantpeakswhichdistributedat:150,700and6000nm,respectively.ThisresultsuggeststhatthesupernatantinMBR-CwasmoreheterogeneousthanthoseinMBR-AandMBR-B.Theheterogeneoussystemcanresultinacomplexmembranefoulingduetothecomplexinteractionbetweentheseparticlesandsolutes.Asynergisticfoulingbehaviorwasfoundduringmembrane ltrationofcolloidalmaterialsanddissolvedmat-ters[37].Thesynergisticfoulingbehaviorisattributedtothehinderedbackdiffusionoffoulantscausedbytheinteractionsbetweenorganicandcolloidalfoulants,whichresultinfasterandmoresubstantialfoulantdepositiononthemembranesurface[37].

Thepeaksat60–800nmwereduetothepresenceofcol-loidsandsolutes,whichcausedbythereleaseofboundEPSfromsludge octosludgesuspension.But,thepeakat6000nmindicatedthepresenceofnon-settleablecellsor ocfragmentsinthesludgesuspension,whichfurthersuggeststhatthe

high

Fig.3.ParticlesizedistributionsofthesludgesupernatantinthethreeMBRs:(a)MBR-Awithanaerationintensityof150L/h,(b)MBR-Bwithanaerationintensityof400L/hand(c)MBR-Cwithanaerationintensityof800L/h.

96F.Mengetal./SeparationandPuri cationTechnology

59 (2008) 91–100

Fig.4.Evolutionof(a)EPSconcentration,(b)colloidalCODand(c)solubleCODduringlong-termmembrane ltrationtests.

aerationintensitycouldleadtoseverebreakageofthesludge ocs.

3.2.3.ChangeofEPS,colloidalCODandsolubleCOD

Inthiswork,thesumoftotalproteinsandcarbohydrateswasconsideredtorepresentthetotalamountofEPSbecausethesearethedominantcomponentstypicallyfoundinextractedEPS[38].ResultsfromsomerecentstudiesindicatethatmainlyproteinandcarbohydrateintheEPScontributetothedeclineofthepermeate ux[39,40].ItwasthusexpectedthatthequantityofEPSwouldcorrelatetomembranefouling.TheEPSconcentrationofMBR-AwassmallerthanthatofMBR-BandMBR-C,indicatingthatthehigheraerationintensitycausedthereleaseoftoomuchEPS(Fig.4a).Duringthewholeexperiment,theEPSconcentrationsofMBR-BandMBR-C

increaseddramatically,thendecreasedandreachedsteadyval-ues.

Ithasbeenobservedthatcolloidalparticlesinthesludgesus-pensionhaveparticularimpactonmembranefouling[41,42].Thesolubleproductscanbereadilydepositedontothemem-branesurfacesbypermeationdrag,andnotreadilydetachedbyshearforceduetoitslowbacktransportvelocity[43].Theevo-lutionofthecolloidalCODofsludgesuspensionisgiveninFig.4b.Wenotethatatthebeginningofthetest,thecolloidalCODofthethreeMBRsincreaseddramatically,andhadhighvaluesfrom100to300h.EventhoughthecolloidalCODofthethreeMBRsdecreasedslowlyandreachedasteadyvalueafterabout400h,thecolloidalCODofMBR-CandMBR-BhavehighervaluesthanthatofMBR-A.Inaddition,asitcanbeseenfromFig.4c,thereisasimilarchangingtendencyofsolubleCODforMBR-A,MBR-BandMBR-C.

Theseresultsindicatethattheintensiveshearstressledto ocbreakageandcausedanincreaseofcolloidalparticlesandsolutesinsludgesuspension.Becausethecolloidalparticleandsoluteshavesmallersize,theycouldresultinaseveremembranefouling.FromFigs.2aand4,wecanseethatEPS,colloidalCODandsolubleCODmayhavesigni cantrelationwithmembranepermeation,thatiswhythepermeatebehaviorofthethreeMBRscouldbeseparatedintotwophases(Fig.2a).

Moreover,theDOconcentrationinducedbyaerationwouldhavesomeeffectonbiomasscharacteristics.Inactivatedsludgeprocess,iftheDOconcentrationistoolow(<2.0mg/L),itcanresultinsludgebulkingbecauseoftheovergrowthof lamen-tousbacteria.But,theDOconcentrationsofthethreeMBRswerechangedfrom3.2to4.76mg/L,andto6.50mg/L,respec-tively.ItindicatesthattherewereenoughDOinthethreeMBRs;therefore,thein uenceofDOconcentrationonsludgepropertycanbeignored.Additionally,theMLSSconcentrations(datawasnotshown)ofthethreeMBRshadlittlechangeinthewholelong-term ltrationtests,sotheimpactofMLSSconcentrationonmembranefoulingalsocanbeignored.3.3.Analysisofmembranefoulants

3.3.1.Deposition/adsorptionmechanismofmembranefoulants

Asthelong-termoperationwasterminated,thethreemem-branemodulesweretakenoutfromthebioreactorsand ushedbypurewater.Thesuspendedsolids(SS),colloidalCOD(CODc)andsolubleCOD(CODs)ofthewashedliquidwereevaluatedtoquantifythefoulantsthataccumulatedonthemem-branesurface(Table2).Atthesametime,thecomponentsofthesludgesuspensionineachMBRweremeasuredtointer-pretthedeposition/adsorptionmechanismofmembranefoulants(Table2).Evidently,themembranefoulantsandsludgesuspen-sionconsistedofsludgeparticles,colloidalparticlesandsolutes.ItcanbeseenthattheSS,whichmainlyconsistedofsludge ocs,decreasedsigni cantlywithincreasingaerationintensity,indi-catingthattheshearstressinducedbyaerationhasgreateffectonthedepositionoflargeparticles.Astheaerationintensityincreasedfrom150to400L/h,therelativecontentofcolloidalparticlesdecreasedfrom8.05%to5.38%,however,therelative

F.Mengetal./SeparationandPuri cationTechnology 59 (2008) 91–100

Table2

Analysisresultsofthemembranefoulants

Componentsofmembranefoulants(%)SS(g/m2)

MBR-AMBR-BMBR-C

34.6(87.31)22.6(88.11)7.9(69.66)

CODc(g/m2)3.19(8.05)1.38(5.38)1.11(9.79)

CODs(g/m2)1.84(4.64)1.67(6.51)2.33(20.55)

Totalfoulants(g/m2)39.6325.6511.34

Componentsofsludgesuspension(%)SS(g/L)6.24(99.54)6.43(99.44)6.59(99.23)

CODc(g/L)0.015(0.24)0.019(0.29)0.027(0.41)

CODs(g/L)0.014(0.22)0.017(0.26)0.024(0.36)

97

Total(g/L)6.276.476.64

contentofsolubleproductsincreasedfrom4.64%to6.51%.Italsocanbeseenthattherelativecontentofcolloidalparti-clesandsolutesonmembranesurfaceweremuchhigherthantheirrelativecontentinthesludgesuspension,con rmingthatthecolloidsandsoluteshaveastrongertendencytodepositontothemembranesurfacethanthesludge ocs.Table2alsodepictsthatinMBR-CthecontributionofcolloidalparticlesandsolutestothetotalfoulantswaslargerthanthatinMBR-AandMBR-B,implyingthatastheMBRoperatedunderhighaera-tionintensity,colloidalparticlesandsolubleproductsmaybethemajorfoulantstomembranefouling.Accordingtoprevi-ousliteratures[43–45],ithasbeenshownthatthecolloidsanddissolvedorganicmatterareresponsibleformembranefoul-ingduringmembrane ltrationofsludgesuspension.Therefore,thecolloidsandsolutesinsludgesuspensionshouldbecon-trolledduringthelong-termoperationofMBRsinordertoobtainpreferablemembranepermeation.

Inthemembrane ltrationprocessunderaconstantTMP,the ltrationprocesscanbedividedintothefollowingtwostages:accumulationstageofundetachablecakelayer,andasecondstagewhereaccumulationanddetachmentofcakelayerreachequilibriumstage[46].InertialliftisthedominantmechanismforlargeparticlesandhighshearrateswhereasBrowniandif-fusionisthedominantforsmallparticlesandlowshearrates[47,48].However,shear-induceddiffusionseemstobethemostimportantoneforintermediateparticlesizesandshearrates[47,48].FromTable2,itcanbeseenthatastheaerationintensityincreasedfrom150to400L/h,thedepositionofcolloidsonthemembranesurfacedecreasedfrom3.19to1.3g/m2,however,thedepositionofsoluteshadlittlevariation.ItisclearthattheBrowniandiffusionwasthemainbacktransportmechanismforMBR-A,butBrowniandiffusionandshear-induceddiffusioncoexistedinMBR-B.Astheaerationintensityincreasedfrom400to800L/hinMBR-C,theconcentrationofthelargeparti-clesdecreaseddramatically.Thedataobtainedfromthecurrentinvestigation,togetherwithpreviousworkintheliterature,con- rmthatthebacktransportmechanismforMBR-CconsistedofBrowniandiffusion,shear-induceddiffusionandinertiallift.

Table3

Analysisresultsof ltrationresistance

Items(%)Rm(1011m 1)

MBR-AMBR-BMBR-C

1.05(6.02)1.05(10.76)1.05(9.44)

Rp(1011m 1)2.45(14.06)2.71(27.77)4.18(37.59)

FromTable2,itcanbeseenthatthesolubleproductsandcolloidalparticleshadagreatcontributiontothecakelayerastheMBRoperatedunderhighshearforcecondition.Itindicatesthatlargeaerationintensitycaninducetheformationofanon-porouscakelayer.Thisisthereasonwhythepermeate uxofMBR-CdecreasedmoreabruptlythanthatofMBR-AandMBR-B.InMBR-A,thelowershearforcecouldnotremovethefoulinglayereffectively,andhenceresultedintheformationofathickerfoulinglayer(seeTable3)onthemembranesurfacewhichwouldincreasethemembrane ltrationresistancestrongly.

3.3.2.Evaluationof ltrationresistance

Toexaminethefoulingtendencies,cakemass,speci ccakeresistanceandeachresistancetermwereanalyzed(Table3).DuringtheoperationofMBRs,sludge ocs,colloidsandsolutesdepositedonthemembranesurface,thecakeresistancebecamethedominantresistance.Thecontributionofcakeresistancetototalresistancehadarangefrom52.97%to79.40%.Table3alsoshowsthatthecakeresistanceinMBR-AwasmorethantwotimesofthoseinMBR-BandMBR-C,indicatingthataera-tionhadgreatimpactsontheremovalofcakelayer.Furthermore,theporeblockingresistanceincreasedwithincreasingaerationintensity,especiallyforMBR-C.ThisresultsuggeststhatthereoccurredasevereirreversiblefoulinginMBR-C.Thisresult,togetherwithTable2,showsthatunderhighaerationinten-sitythedepositionandadsorptionofcolloidsandsolutesonmembranewouldresultinsevereporeblockingorirreversiblefouling.

Theeffectofaerationoncakemassandspeci cresistanceisshowninTable3.Thecakemassdecreaseddramaticallyasaer-ationintensityincreased,however,thespeci ccakeresistanceincreasedde nitelyastheaerationintensityincreasedfrom400to800L/h.Itindicatesthatthedepositionofcolloidsandsolutesonmembranesurfacewouldformadensecakelayer.AccordingtoCarmanKozenyequation,thespeci cresistanceissigni -cantlyin uencedbysludgeparticlesizeandcakeporosity.InMBR-C,thecolloidsandsolutesweresigni cantcontributorstothefoulantsthatdepositedonthemembranesurface,soitcouldformadensecakelayer.

Rc(1011m 1)13.84(79.40)6.00(61.48)5.89(52.97)

Rt(1011m 1)17.439.7611.12

mc(g/m2)39.6325.6511.34

α(1010m/g)3.492.345.19

98F.Mengetal./SeparationandPuri cationTechnology

59 (2008) 91–100

Fig.5.FTIRspectraofthemembranefoulants:(a)MBR-A,(b)MBR-Band(c)MBR-C.

F.Mengetal./SeparationandPuri cationTechnology 59 (2008) 91–10099

3.3.3.FTIRanalysisofmembranefoulants

Ingeneral,theFTIRtechniquecanprovidemoredetailedinformationaboutthedepositionofbiopolymersonthemem-branesurface.TheFTIRspectraofmembranefoulantsinthethreeMBRsarepresentedinFig.5.Theyaresimilarinthepro- lebutsigni cantlydifferentintheadsorptionintensity.Thespectrumshowsabroadregionofabsorptionat3400cm 1,whichisduetothestretchingoftheO–Hbondinhydroxylfunc-tionalgroups[49].Therewasabroadpeakat1100cm 1,whichisduetoC–Obondsandisassociatedwithalcohols,ethersandpolysaccharides.InMBRs,thispeakisusuallyattributedtothepresenceofpolysaccharidesorpolysaccharide-likesub-stances.Choetal.[50]attributedthispeaktopolysaccharidesorpolysaccharide-likemembranefoulants;whereasThurman[51]attributedthispeaktosilicateimpuritiesinhumicsamples.AsshowninFig.5,therearetwopeaks(1640and1550cm 1)inthespectrumwhichareuniquetotheproteinsecondarystructure,calledamidesIandII[52].TheamideIisthestretchingvibrationbandsassociatedprimarilywiththepeptidecarbonyls(CO),andtheamideIIbandsat1550cm 1isduetotheinteractionbetweentheN–HbondingandtheC–NstretchingoftheC–N–Hgroup[53].Thisresultindicatesthepresenceofproteinsinmem-branefoulants.Basedonthepeakat1380cm 1,themembranefoulantscontainedamediumamountoflipids[54].Jarusutthiraketal.observedasigni cantpeakat1720cm 1duringmembrane ltrationofwastewatertreatmentplantef uent[55].Thispeakisassociatedwithcarboxylicgroups,representingatypicalchar-acteristicsofhumicandfulvicacids.Inourwork,thispeakwasabsentindicatingthattherewasnohumicorfulvicacidsinthemembranefoulantsortheamountofhumicandfulvicacidsinthemembranefoulantscouldbeignored.Thepresenceofproteins,polysaccharidesandlipidsinmembranefoulantssug-gestsasigni cantorganicfoulingwhichmainlyresultedfromEPS.

FromFig.5,italsocanbeseenthattheabsorptioninten-sityofFTIRspectrumsinthethreeMBRswasdifferentfromeachother.Theintensitywasgivenintheorderof:MBR-C>MBR-B>MBR-A.Theabsorptionintensityre ectedtherelativeamountofbiopolymersinthetotalfoulants.Therefore,theabove-mentionedresultindicatesthattherelativeamountofbiopolymersorEPSinthetotalfoulantsalsofollowedtheorder:MBR-C>MBR-B>MBR-A.ThisresultcoincidedwiththeresultsobtainedfromSection3.3.1.4.Summary

Thispaperpresentsacomparativeandcorrelativestudyofaerationintensitiesonmembraneperformance.ThemembranefoulingmechanismsofthreeparallelMBRswereinvestigatedfromtwoaspects:analysisofsludgecharacteristicsandevalua-tionofmembranefoulants.Fromtheresultsreportedhere,thefollowingconclusionscanbedrawn:

Aerationintensityhadsigni cantimpactsonmembraneper-meation.Smallorlargeaerationintensityhadanegativein uenceonmembranepermeability.Lowaerationcouldnotremovethemembranefoulantsfrommembranesurfaceeffec-

tively.However,highaerationcouldinduceaseverebreakageofsludge ocs.

Thehighaerationintensitygeneratedastrongershearstress,andthenresultedinasevere ocbreakage.Thesmallsizeofsludgeparticlesgeneratedatthehighershearstresswasincloseassociationwiththedramaticmembranefouling.Furthermore,thehighshearconditioncouldgenerateahet-erogeneoussludgesupernatant,andresultedinacomplexmembranefouling.Thehighshearconditionscouldpromotethereleaseofcolloidalandsolublecomponentsfromthemicrobial ocstothebulksolutionduetomicrobial ocbreakageandthuscausearapidlossinmembraneperme-ability.

Themembranefoulantsconsistedofsludge ocs,colloidalparticlesandsolutes.ThecontributionofcolloidalparticlesandsolutestothemembranefoulantsbecamemoreimportantastheMBRoperatedunderlargeaerationintensity.ThethreeMBRshaddifferentbacktransportmechanismoffoulantsdeposition.Underaerationintensityof150L/h,Browniandif-fusionwasthemainbacktransportmechanismformembranefoulants,whichcouldnotremovalthecakelayereffectively.ThecakeresistanceofMBR-A(150L/h)wasmorethantwotimesofMBR-B(400L/h)andMBR-C(800L/h),indicatingthataerationhasgreatimpactsontheremovalofcakelayer.Thehighaerationintensity(800L/h)couldresultinseveremembraneporeblocking.WiththehelpofFTIRtechnique,themajorcomponentsoforganicmattersinthemembranefoulantswereidenti edasproteins,polysaccharidemattersandlipids.Acknowledgement

TheprojectsupportedbyNationalNaturalScienceFounda-tionofChina,GrantNo.50578024.References

[1]H.Y.Ng,T.W.Tan,S.L.Ong,Membranefoulingofsubmergedmembrane

bioreactors:impactofmeancellresidencetimeandthecontributingfactors,Environ.Sci.Technol.40(2006)2706–2713.

[2]K.Kimura,N.Yamato,H.Yamamura,Y.Watanabe,Membranefoulingin

pilot-scalemembranebioreactors(MBRs)treatingmunicipalwastewater,Environ.Sci.Technol.39(2005)6293–6299.

[3]H.Choi,K.Zhang,D.D.Dionysiou,D.B.Oerther,G.A.Sorial,Effectof

permeate uxandtangential owonmembranefoulingforwastewatertreatment,Sep.Purif.Technol.45(2005)68–78.

[4]X.-M.Wang,X.-Y.Li,X.Huang,Membranefoulinginasubmergedmem-branebioreactor(SMBR):characterisationofthesludgecakeanditshigh ltrationresistance,Sep.Purif.Technol.52(2007)439–445.

[5]S.Ognier,C.Wisniewski,A.Grasmick,Membranebioreactorfoulingin

sub-critical ltrationconditions:alocalcritical uxconcept,J.Membr.Sci.229(2004)171–177.

[6]J.Orantes,C.Wisniewski,M.Heran,A.Grasmick,Thein uenceof

operatingconditionsonpermeabilitychangesinasubmergedmembranebioreactor,Sep.Purif.Technol.52(2006)60–66.

[7]L.Chu,S.Li,Filtrationcapabilityandoperationalcharacteristicsof

dynamicmembranebioreactorformunicipalwastewatertreatment,Sep.Purif.Technol.51(2006)173–179.

[8]W.Khongnakorn,C.Wisniewski,L.Pottier,L.Vachoud,Physicalproper-tiesofactivatedsludgeinasubmergedmembranebioreactorandrelationwithmembranefouling,Sep.Purif.Technol.55(2007)125–131.

100F.Mengetal./SeparationandPuri cationTechnology 59 (2008) 91–100

[9]M.Brockmann,C.F.Seyfried,Sludgeactivityundertheconditionsof

cross owmicro ltration,WaterSci.Technol.35(1997)173–181.

[10]K.Yamamoto,M.Hiasa,T.Mahmood,T.Matsuo,Directsolid–liquidsep-arationusinghollow bermembraneinanactivatedsludgeaerationtank,WaterSci.Technol.21(1989)43–54.

[11]S.Churchouse,Membranebioreactors:goingfromlaboratorytolarge

scale-problemstoclearsolutions,in:ProceedinginMembranesandtheEnvironment,UniversityofOxford,2002.

[12]T.Ueda,K.Hata,Y.Kikuoka,O.Seino,Effectsofaerationonsuctionpres-sureinasubmergedmembranebioreactor,WaterRes.31(1997)489–494.[13]S.P.Hong,T.H.Bae,T.M.Tak,S.Hong,A.Randall,Foulingcontrolinacti-vatedsludgesubmergedhollow bermembranebioreactors,Desalination143(2002)219–228.

[14]S.-S.Han,T.-H.Bae,G.-G.Jang,T.-M.Tak,In uenceofsludgereten-tiontimeonmembranefoulingandbioactivitiesinmembranebioreactorsystem,Proc.Biochem.40(2005)2393–2400.

[15]T.-H.Bae,S.-S.Han,T.-M.Tak,Membranesequencingbatchreactorsys-temforthetreatmentofdairyindustrywastewater,Proc.Biochem.39(2003)221–231.

[16]L.Defrance,M.Y.Jaffrin,Comparisonbetween ltrationsat xedtrans-membranepressureand xedpermeate ux:applicationtoamembranebioreactorusedforwastewatertreatment,J.Membr.Sci.152(1999)203–210.

[17]T.Ueda,K.Hata,Domesticwastewatertreatmentbyasubmerged

membranebioreactorwithgravitational ltration,WaterRes.33(1999)2888–2892.

[18]X.Zheng,J.Liu,Dyeingandprintingwastewatertreatmentusingamem-branebioreactorwithagravitydrain,Desalination190(2006)277–286.[19]X.Zheng,J.X.Liu,Optimizingofoperationalfactorsofamembrance

bioreactorwithgravitydrain,WaterSci.Technol.52(2005)409–416.[20]B.Fan,X.Huang,Characteristicsofaself-formingdynamicmembrane

coupledwithabioreactorformunicipalwastewatertreatment,Environ.Sci.Technol.36(2002)5245–5251.

[21]B.Fround,R.Palmgren,K.Keiding,P.H.Nielsen,Extractionofextracel-lularpolymersfromactivatedsludgeusingacationexchangeresin,WaterRes.30(1996)1749–1758.

[22]O.H.Lowery,N.J.Rosebrough,A.L.Farr,R.J.Randall,Proteinmeasure-mentwiththefolinphenolreagent,J.Biol.Chem.193(1951)265–275.[23]J.-S.Park,K.-M.Yeon,C.-H.Lee,Hydrodynamicsandmicrobial

physiologyaffectingperformanceofanewMBR,membrane-coupledhigh-performancecompactreactor,Desalination172(2005)181–188.

[24]J.Lee,W.Y.Ahn,C.H.Lee,Comparisonofthe ltrationcharacteristics

betweenattachedandsuspendedgrowthmicroorganismsinsubmergedmembranebioreactor,WaterRes.35(2001)2435–2445.

[25]F.G.Meng,H.M.Zhang,Y.S.Li,X.W.Zhang,F.L.Yang,J.N.Xiao,Cake

layermorphologyinmicro ltrationofactivatedsludgewastewaterbasedonfractalanalysis,Sep.Purif.Technol.44(2005)250–257.

[26]F.G.Meng,H.M.Zhang,Y.S.Li,X.W.Zhang,F.L.Yang,Applicationof

fractalpermeationmodeltoinvestigatemembranefoulinginmembranebioreactor,J.Membr.Sci.262(2005)107–116.

[27]APHA,StandardMethodsForTheExaminationofWaterandWastewater,

19thed.,AmericanPublicHealthAssociation,Baltimore,MD,1995.[28]G.Belfort,R.H.Davis,A.L.Zydney,Thebehaviorofsuspensionsand

macromolecularsolutionsincross owmicro ltration,J.Membr.Sci.96(1994)1–58.

[29]B.J.James,Y.Jing,X.DongChen,Membranefoulingduring ltrationof

milk—amicrostructuralstudy,J.FoodEng.60(2003)431–437.

[30]K.Yamagiwa,Y.Oohira,A.Ohkawa,Performanceevaluationofaplunging

liquidjetbioreactorwithcross ow ltrationforsmall-scaletreatmentofdomesticwastewater,Bioresour.Technol.50(1994)131–138.

[31]F.S.Fan,H.D.Zhou,H.Husain,Identi cationofwastewatersludgechar-acteristicstopredictcritical uxformembranebioreactorprocesses,WaterRes.40(2006)205–212.

[32]J.Zhang,H.C.Chua,J.Zhou,A.G.Fane,Factorsaffectingthemem-braneperformanceinsubmergedmembranebioreactors,J.Membr.Sci.284(2006)54–66.

[33]A.Drews,M.Vocks,V.Iversen,B.Lesjean,M.Kraume,In uenceof

unsteadymembranebioreactoroperationonEPSformationand ltrationresistance,Desalination192(2006)1–9.

[34]P.Le-Clech,V.Chen,T.A.G.Fane,Foulinginmembranebioreactorsused

inwastewatertreatment,J.Membr.Sci.284(2006)17–53.

[35]R.Bai,H.F.Leow,Micro ltrationofactivatedsludgewastewater—the

effectofsystemoperationparameters,Sep.Purif.Technol.29(2002)189–198.

[36]D.S.Parker,W.J.Kaufman,D.Jenkins,Flocbreakupinturbulent occu-lationprocesses,J.Sanit.Eng.Div.ASCESA1(1972)79–89.

[37]Q.Li,M.Elimelech,Synergisticeffectsincombinedfoulingofaloose

nano ltrationmembranebycolloidalmaterialsandnaturalorganicmatter,J.Membr.Sci.278(2006)72–82.

[38]W.Lee,S.Kang,H.Shin,Sludgecharacteristicsandtheircontributionto

micro ltrationinsubmergedmembranebioreactors,J.Membr.Sci.216(2003)217–227.

[39]Y.Ye,P.L.Clech,V.Chen,A.G.Fane,Evolutionoffoulingduringcross- ow ltrationofmodelEPSsolutions,J.Membr.Sci.264(2005)190–199.

[40]H.Nagaoka,H.Nemoto,In uenceofextracellularpolymericsubstanceon

nitrogenremovalinanintermittently-aeratedmembranebioreactor,WaterSci.Technol.51(2005)151–158.

[41]E.H.Bouhabila,R.BenAim,H.Buisson,Foulingcharacterisation

inmembranebioreactors,Sep.Purif.Technol.22–23(2001)123–132.

[42]L.Defrance,M.Y.Jaffrin,B.Gupta,P.Paullier,V.Geaugey,Contribution

ofvariousconstituentsofactivatedsludgetomembranebioreactorfouling,Bioresour.Technol.73(2000)105–112.

[43]T.H.Bae,T.M.Tak,Interpretationoffoulingcharacteristicsofultra ltration

membranesduringthe ltrationofmembranebioreactormixedliquor,J.Membr.Sci.264(2005)151–160.

[44]S.Rosenberger,abs,B.Lesjean,R.Gnirss,G.Amy,M.Jekel,J.C.

Schrotter,Impactofcolloidalandsolubleorganicmaterialonmembraneperformanceinmembranebioreactorsformunicipalwastewatertreatment,WaterRes.40(2006)710–720.

[45]S.Rosenberger,H.Evenblij,S.tePoele,T.Wintgens,abs,Theimpor-tanceofliquidphaseanalysestounderstandfoulinginmembraneassistedactivatedsludgeprocesses—sixcasestudiesofdifferentEuropeanresearchgroups,J.Membr.Sci.263(2005)113–126.

[46]R.Bian,K.Yamamoto,Y.Watanabe,Theeffectofshearrateoncontrolling

theconcentrationpolarizationandmembranefouling,Desalination131(2000)225–236.

[47]C.Wisniewski,A.Grasmick,A.LeonCruz,Criticalparticlesizeinmem-branebioreactors:caseofadenitrifyingbacterialsuspension,J.Membr.Sci.178(2000)141–150.

[48]R.H.Davis,Modelingoffoulingofcross owmicro ltrationmembranes,

Sep.Purif.Methods21(1992)75–126.

[49]M.Kumar,S.S.Adham,W.R.Pearce,Investigationofseawaterreverse

osmosisfoulinganditsrelationshiptopretreatmenttype,Environ.Sci.Technol.40(2006)2037–2044.

[50]J.Cho,G.Amy,J.Pellegrino,Y.Yoon,Characterizationofcleanandnat-uralorganicmatter(NOM)fouledNFandUFmembranes,andfoulantscharacterization,Desalination118(1998)101–108.

[51]E.M.Thurman,OrganicGeochemistryofNaturalWaters,Martinus

Nijhoff/Dr.W.JunkPublishers,Boston,MA,1985.

[52]T.Maruyama,S.Katoh,M.Nakajima,H.Nabetani,T.P.Abbott,A.Shono,

K.Satoh,FT-IRanalysisofBSAfouledonultra ltrationandmicro ltrationmembranes,J.Membr.Sci.192(2001)201–207.

[53]R.M.Silverstein,F.X.Webster,SpectrometricIdenti cationofOrganic

Compounds,sixthed.,Wiley,NewYork,1998.

[54]A.Ramesh,D.J.Lee,i,Membranebiofoulingbyextracellularpoly-mericsubstancesorsolublemicrobialproductsfrommembranebioreactorsludge,Appl.Microbiol.Biotechnol.74(2007)699–707.

[55]C.Jarusutthirak,G.Amy,J.-P.Croue,Foulingcharacteristicsofwastew-ateref uentorganicmatter(EfOM)isolatesonNFandUFmembranes,Desalination145(2002)247–255.

本文来源：https://www.bwwdw.com/article/k1oe.html